Plasma levels of bone Gla-.protein reflect bone formation in patients on chronic maintenance dialysis

Transcription

1 Kidney International, Vol 26 (1984), pp Plasma levels of bone Gla-protein reflect bone formation in patients on chronic maintenance dialysis HARTMUT H MALLUCHE, MARIE-CLAUDE FAUGERE, PAOLO FANTI, and PAUL A PRICE Division of Nephrology, Bone and Mineral Metabolism, Department of Medicine, University of Kentucky, Lexington, Kentucky, and Department of Biology, University of Californi an Diego, California Plasma levels of bone Gla-protein reflect bone formation in patients on chronic maintenance dialysis Predictive value of plasma levels of bone Gla-protein (BGP) for bone histology was evaluated in 3 chronically dialyzed patients All patients underwent bone biopsies and serum biochemical parameters, including BUP, parathyroid hormone, and alkaline phosphatase; calcium and phosphate were measured at the time of biopsy Bone histology showed renal osteodystrophy with low bone turnover and osteomalacia (LT-ROD) in 13 patients, and renal osteodystrophy with high bone turnover and prevailing hyperparathyroid bone disease (HT-ROD) in 17 patients Values for BGP were above normal in LT-ROD ( vs 68 2 ng/ml) and extremely elevated in HT-ROD ( ng/ml) imilar differences were not found with the other serum biochemical parameters, even though BGP correlated with parathyroid hormone (r = 64) and alkaline phosphatase (r = 85) There were significant correlations between BGP and cellular and non-cellular parameters of bone formation (r = 73 to 91) Weaker or no correlations were found between BGP and histologic parameters of bone, reflecting mainly mineralization or resorption These correlations and the finding of significant differences in plasma BGP between LT-ROD and HT-ROD indicate that plasma levels of BGP reflect bone formation in uremia and predict underlying bone histology Les taux plasmatiques de Ia Gal-proteine osseuse reflètent Ia formation osseuse chez les malades en dialyse chronique La valeur des taux plasmatiques de Ia Gla-protcine osseuse pour Ia prediction de l'histologie osseuse a été ëvaluée chez 3 hemodialyzes chroniques Une biopsie osseuse a été pratiquée chez tous les malades et les prélèvements sanguins pour Ia mesure de Ia BGP, parathormone, alcaline phosphatase, calcium et phosphore ont été effectucs au moment de Ia biopsie L'histologie osseuse a révclé une ostéodystrophie rénale a faible niveau de remodelage (LT-ROD) chez 13 malades et une ostéodystrophie renal a haut niveau de remodelage avec hyperparathyroidisme (HT-ROD) chez 17 malades Les taux de BGP étaient au dessus de la normale chez les malades avec LT-ROD ( vs 68 2ng/ml) et extrcmement Clevés chez les malades avec HT-ROD ( ng/ml) Aucune difference entre les deux groupes n'a été retrouvce avec Ia parathormone ou alcaline phosphatase hien que BGP soit correlée avec Ia parathormone (r = 64) et Ia phosphatase alcaline (r = 85) Des correlations significatives ont etc établies entre Ia BGP et les paramctres cellulaires et non cellulaires de Ia formation osseuse (r 73 a 91) Des correlations plus faibles ou une absence de correlation ont etc trouvces entre BGP et les paramètres histologiques osseux de La minéralisation et de Ia resorption Ces correlations et Ia presence de differences significatives des taux plasmatiques de la BGP entre LT-ROD et HT-ROD indiquent que Ia BGP plasmatique refictent Ia formation osseuse chez les malades urémiques et predisent les anomalies histologiques osseuses reflects its content of three residues of the vitamin K-dependent amino acid, y-carboxyglutamic acid (GLA) BGP in plasma has the same apparent molecular weight as the pure bone GLAprotein and studies indicate that the plasma protein is probably the intact bone protein [31 With the availability of a radioimmunoassay for detection of BGP in plasma [3], it became possible to study BGP levels in various pathologic situations [4 7] Because bone and kidney may affect blood concentrations of BGP [8, 9], patients without kidney function allowed us to study the contribution of bone to blood levels of BGP rrelations between BGP and other serum biochemical parameters as well as histomorphometric parameters of bone in these patients might help us understand the role of BGP in bone metabolism and/or its value for predicting underlying bone histology In the present study we measured plasma BGP levels and other serum biochemical parameters in patients on chronic maintenance dialysis and evaluated the relationship between plasma BGP and quantitative static and dynamic parameters of bone structure, bone formation and resorption In addition, we studied the relative values of plasma BGP, serum levels of alkaline phosphatase, and parathyroid hormone (PTH) for the prediction of histologic changes in patients with uremic osteodystrophy Methods Patients Thirty consecutive patients who were scheduled for diagnostic bone biopsies were enrolled in the study There were 18 male patients and 12 female patients with a mean age of years (range, 13 to 72 years) All patients were on chronic maintenance dialysis Twenty-six of them were on hemodialysis and four on chronic ambulatory peritoneal dialysis Informed consent was obtained from all patients before the study The duration of dialytic therapy was months (range, 2 to 18 months) All patients were anuric None of the patients were parathyroidectomized or nephrectomized and none of them had been immobilized No other medical problems such as liver cirrhosis, diabetes mellitus, or malabsorption were known to exist All patients had normal outdoor activities and were on Bone Gla-protein (BGP), also referred to as 'osteocalcin," represents one of the most abundant non-collagenous bone proteins [1, 2] It has a molecular weight of 58 and the name Received for publication February 3, 1984, and in revised form June 7, by the International ociety of Nephrology 869

2 87 Malluche ci a! Table I tatic and dynamic quantitative parameters of bone structure, bone formation, and bone resorption in 3 dialyzed uremic patients" Parameters of bone structure Bone mass, % Lamellar osteoid volume, mm3/cm3 Woven osteoid volume, mm3icm Parameters of bone formation and resorption Lamellar osteoid surface, % Woven osteoid surface, % Thickness of lamellar osteoid, pm Bone-osteoblast interface, % Peritrabecular fibrosis, % Bone-osteoclast interface, % Parameters of bone dynamics Mineral apposition rate, pm/day Labelled osteoid seams, % Mineralization lag time, days LT-ROD 181 1W' HT-ROD Normal ntrols 25 39" l Abbreviations: LT-ROD, renal osteodystrophy with low bone turnover and osteomalacia; HT-ROD, renal osteodystrophy with high bone turnover and hyperparathyroid bone disease "Values are given as mean E b Normal values obtained from 84 normal American subjects [12]; values are given as mean D Normal values for dynamic parameters were obtained from 28 normal American subjects an unrestricted diet except for potassium and fluids The hemodialyzed patients were dialyzed three times weekly, 5 hr each, using the hollow fiber dialyzer (rdis Dow rp, Miami, Florida) with 15 m2 surface and dialysis machines (model AK1O, Travenol Laboratories, Inc, Morton Grove, Illinois) The dialysate contained 35 meq calcium and 2 meq magnesium No concurrent therapy was given with the exception of routine dialysis support medications including folic acid, iron, vitamin B6, phosphate binders, and multivitamins Bone biopsies Biopsy specimens were taken from the anterior iliac crest using an electric drill [11 Tetracycline hydrochloride was given (Tetracyn, 5 mg po bid) on day 19 and 2 before the biopsies, The drug was stopped for the following 12 days, and subsequently, demeclocycline hydrochloride (Declomycin 3 mg po bid) was given for 4 days Bone biopsy specimens were obtained 4 days thereafter Bone histology and histomorphometry Bone specimens were fixed in ethanol for 24 hr, dehydrated, and embedded in methylmethacrylate Undecalcified sections of 3- and 7- thickness were prepared using a microtome (Model 114, Jung, Heidelberg, Germany) The sections were stained with the modified Goidner's trichrome stain [111, which permits discrimination of calcified bone from osteoid and gives excellent cellular details [121, even-micron thick unstained sections were prepared for phase-contrast and fluorescence microscopy In addition, all slides were stained with a specific histochemical stain for detection of aluminum in bone [131 The slides were read without knowledge of clinical or biochemical information All sections were analyzed quantitatively for static and dynamic parameters of bone structure, bone formation, and resorption using the Osteoplan (Carl Zeiss, Thornwood, New York) according to Malluche et al [141 A minimum of 5 optical fields were evaluated at a magnification of x 2 using an objective with 4 numerical aperture Radioimmunoassay for BGP in plasma and dialysate At time of bone biopsies, blood samples were taken from all 3 patients for determination of plasma concentrations of BGP BGP was measured without knowledge of bone histology in triplicate using a radioimmunoassay with rabbit antibody directed against calf BGP [31 The antibody crossreacts with purified human BGP but not with BGP from rat or rabbit bone tudies with peptides of known structure derived from enzymatic digestion of BGP indicate that the rabbit antibody recognizes the COOH-terminal region of the 49-residue calf bone protein All determinations were done with four different dilutions (1, 1, 1, and 1 jd) to ascertain agreement between different dilutions and to assure that results were on scale The detection limit of the assay is 1 ng, the intraassay variation is less than 1%, and the intcrassay variation is less than 15% [3] To determine the effect of dialysis on plasma BGP levels, we measured plasma BGP in 6 of the 3 patients at the beginning and the end of a dialysis without ultrafiltration In addition, blood samples were obtained before and after passage through the dialyzer, that is, from arterial and venous sites, and dialysate samples were taken from the outflow of the dialyzer To determine day-to-day variations in plasma BGP levels of patients on chronic maintenance dialysis, we followed prospectively plasma BGP levels in these six patients For this purpose, blood was drawn at least once monthly before and after dialysis for 7 to 9 months Duration and schedule of dialysis were not changed during this time Biochemical measurements erum PTH levels were measured using a sensitive N-terminal parathyroid hormone radioimmunoassay which measures intact PTH (anti-human 1 34 PTH, courtesy of Dr David Endres, Nichols Laboratory, an Juan Capistrano, California) [15] erum concentrations of calcium were measured by atomic absorption spectrophotometry (Perkin-Elmer Model 5, Norwalk, nnecticut) erum phosphate, alkaline phosphatase, and creatinine were measured with an autoanalyzer (model MA 12, Technicon, Tarrytown, New York) tatistical analysis Linear regression analysis, multiple regression model analysis, and correlations were calculated between plasma BGP levels, PTH, alkaline phosphatase, and

3 Plasma BGP levels reflect bone formation 871 Table 2 erum biochemical parameters in 3 dialyzed uremic patients with various types of uremic bone disease' LT-ROD HT-ROD Normal Range CmEq/iiter 47± P, mg/dl Creatinine, mgldi AP, IU ipth, pg/mi BGP, ng/,nl Abbreviations: LT-ROD, Renal osteodystrophy with low bone turnover and osteomalacia; HT-ROD, Renal osteodystrophy with high bone turnover and hyperparathyroid bone disease a Values are given as mean EM 2 Table 3 rrelations (r-values) between histomorphometric parameters of bone and bone Gla-protein (BGP), alkaline phosphatase (AP), and parathyroid hormone (PTH) BGP AP PTH Parameters of bone structure Bone mass Lamellar osteoid volume Woven osteoid volume 73b 69a 62a Parameters of bone formation and resorption Lamellar osteoid surface 6 56 Woven osteoid surface 84" 58a 72b Bone-osteoblast interface 77b 66a 72b Osteoblastic index 78" 68a 73b Peritrabecular fibrosis 91" 76" 84b Bone-osteoclast interface 62a 63a 56a Osteoclastic index ofi b Parameters of bone dynamics Mineral apposition rate 91" 72" 92" Labelled osteoid seams k' Mineralization lag time a p <1 b p <1 15 E '1) - 1 C" E ) C" 5 LT-ROD HT-ROD 1 Plasma levels of bone Gla-protein in patients on chronic Fig maintenance dialysis with low bone turnover and osteomalacia (LT- ROD) and with high bone turnover and hyperparathyroid bone disease (HT-ROD) static and dynamic parameters of bone formation and resorption rrelations were calculated with correction for repetitive sampling Differences in serum biochemical parameters between the two histologic groups of uremic bone disease (see below) were calculated using the Mann-Whitney Rank um test for nonparametric differences All computations were done using the Proc-Gim-In statistical analysis system #979 [16] Results Bone histology Thirteen patients were found to have renal osteodystrophy with low bone turnover and osteomalacia (LT- ROD) The other 17 patients had renal osteodystrophy with high bone turnover and hyperparathyroid bone disease (HT- ROD) Histomorphometric data in these patients are shown in Table 1 Renal osteodystrophy with low bone turnover and osteomalacia was histologically characterized by accumulation of lamellar osteoid, low or normal number of osteoblasts and osteoclasts per unit trabecular surface, low fraction of actively mineralizing (doubly labelled) osteoid seams, decreased mineral apposition rate and prolonged mineralization lag time (Table 1) tainable bone aluminum was seen in 39% of these patients The fraction of bone osteoid interface exhibiting stainable bone aluminum ranged from 62 to 1% The histologic features of renal osteodystrophy with high bone turnover and hyperparathyroid bone disease were high normal or elevated cancelbus bone mass, accumulation of osteoid of abnormal, irregular "woven" collagen structure, abundance of osteoclasts and osteoblasts, and peritrabecular fibrosis The fraction of actively mineralizing osteoid seams was below normal, yet higher than in patients with LT-ROD The mineral apposition rate was high normal or elevated and mineralization lag time was normal tainable bone aluminum was seen in one of these patients with 12% of the osteoid-bone interface exhibiting stainable bone aluminum erum biochemistry There were no significant differences between serum calcium levels of patients with LT-ROD and HT-ROD (Table 2) erum phosphate levels were elevated in both groups, but there was no difference between LT-ROD and HT-ROD erum alkaline phosphatase levels were at the upper normal range in patients with LT-ROD and clearly elevated in patients with HT-ROD erum parathyroid hormone levels were elevated in both groups of patients, and patients with HT-ROD had significantly higher circulating levels of PTH, Plasma BGP Plasma BGP levels were abnormal in all patients Patients with HT-ROD had significantly higher plasma levels of BGP than patients with LT-ROD and there was no overlap in plasma BGP between the two groups (Fig 1) The highest plasma level of BGP in LT-ROD was 121 ng/ml, and the lowest value of plasma BGP in HT-ROD was 278 ng/ml (normal,

4 872 Malluche et al 8 A 6 B 'I) U, a' E (I) y = Iog(BGP) = 86 U, y = BGP 25 x 1-6 (BGP)' r = 84 I I I I I I I I Plasma levels of BGP, log Plasma levels of BGP, ng/ml rrelation between plasma levels of' bone Gin-protein and surface of lamellar osteoid in patients with low bone turnoverand Fig 2 A otleo,naiacia () and with high bone turnover and hyperparalhvroid bone disease ( ) B rrelation between plasma levels of bone Gin-protein and surface of woven osteojd in patients with low hone turnover and osteomcjlacia () and with high bone turnover and hvperparathvroid bone disease ( ) I 68 2 nglml) Plasma BGP levels measured in six patients at the beginning of a dialytic therapy were not different before and after passage through the dialyzer ( vs ng/ml) and no significant changes were found at the beginning and the end of a dialysis ( vs ng/ml) BGP concentrations in dialysate were undetectable in five patients and 34 ng/ml in another patient Plasma BGP levels drawn prospectively in six patients before and after dialysis did not change significantly during 7 to 9 months rrelations between plasma BGP and serum biochemistry There were no correlations between plasma concentrations of BGP and serum levels of calcium, phosphate, and creatinine However, plasma BGP correlated well with serum alkaline phosphatase (r = 85; P < 1) and to a lesser degree with PTH (r = 64; P < 1) rrelations between hone histology, plasma BGP, and serum biochemistry Cancellous bone mass correlated neither with plasma BGP, serum alkaline phosphatase, or PTH Vol ume of lamellar osteoid correlated inversely with BGP but not with alkaline phosphatase or serum PTH There was a direct correlation between volume of woven osteoid, BGP, and PTH The surface of lamellar osteoid correlated negatively with BGP (Fig 2A) and to a lesser degree with alkaline phosphatase and PTH The surface of woven osteoid correlated positively with BGP (Fig 2B) and, again, to a lesser degree with alkaline phosphatase and PTH Other osteoblastic parameters such as bone-osteoblast interface, osteoblastic index, and peritrabecular fibrosis (Fig 3) were also best correlated with plasma BGP and less with alkaline phosphatase and PTH Boneosteoclast interface and osteoclastic index correlated equally with BGP, alkaline phosphatase, and PTH However, these correlations were weaker than those between osteoblastic parameters and BGP Mineral apposition rate correlated well with BGP (Fig 4) and PTH and somewhat less with alkaline phosphatase, whereas actively mineralizing osteoid seams correlated with PTH only Mineralization lag time did not correlate with BGP, alkaline phosphatase, or PTH Discussion The observed correlations between BGP and histomorphometric parameters of bone indicate that BGP is a good marker for osteoblastic activity, in particular for bone matrix and collagen fiber production This is evidenced by a strong correlation between BGP and the volume of woven osteoid, the surface of woven osteoid, and peritrabecular fibrosis The negative correlation between BGP and volume and surface of lamellar osteoid might reflect that, in uremic osteodystrophy, the appearance of woven osteoid is associated with less lamellar osteoid [171 The good correlations between BGP and the number of osteoblasts, bone-osteoblast interface, and mineral apposition rate further support the notion that BGP is an indicator of osteoblastic activity Osteoblastic activity encompasses bone formation and mineralization The lack of correlations between BGP and parameters of bone mineralization, such as fractional labelling of osteoid and mineralization lag time, indicates that BGP reflects mainly bone formation activity by osteoblasts and not bone mineralization in our uremic patients It is of note that BGP levels correlate only marginally with the extent of active resorption lacunae per trabecular surface and

5 Plasma BGP levels reflect bone formation 873 ) ) y = = BGP 35 x 1 (BGP) Plasma levels of BGP, np/rn! Fig 3 rrelation between plasma levels of bone Glaprotein and peritrabecular fibrosis in patients with low bone turnover and osteomalacia () and with high bone turnover and hyperparathyroid bone disease ( ) V = 298 r = BGP 6 x io (BGP) Plasma levels of BGP, np/rn! Fig 4 rrelation between plasma levels of hone Gb-protein and mineral apposition rate in patients with low bone turnover and osteomalacia () and with high bone turnover and hyperparathyrqjd bone disease ( ) the number of osteoclasts per trabecular surface This correlation might be spurious, reflecting the known coupling between bone formation and resorption [181 Our results point to the osteoblasts as the mostlikely site of BGP production This notion agrees with the observation that plasma BGP arises from synthesis of BGP in bone and not from release of BGP through bone resorption [8] In addition, it is in keeping with the finding that BGP is secreted in vitro by osteosarcoma cells which have a PTH responsiveness and alkaline phosphatase activity akin to osteoblasts [19] Our data agree with reports by Delmas et al [6] who found a correlation between plasma BGP levels and bone turnover in osteoporotic patients However, the prevailing notion that BGP is involved in mineralization cannot be confirmed by our results The apparent discrepancy can be explained by the fact that studies ascribing a role fo BGP in mineralization were done either in vitro [2, 211 or in growing rats [22, 231 Our studies were done in patients with various levels of elevated BGP The finding that BGP correlates better with histomorphometric parameters of bone formation than alkaline phosphatase or PTH ascribes a promising role to BGP as a non-invasive tool for assessment of underlying histology This potential is further increased by the fact that BGP, unlike alkaline phosphatase, is a bone-specific protein, that is, not produced by other organs Acknowledgments This study was presented in part at the XVth Annual Meeting of the American ociety of Nephrology, Chicago, 1982, and supported in part by a grant from Dialysis Clinics, Inc (# 129) We thank R Wheaton for skillful technical assistance, and Drs C Wood and R Kryscio, Department of Biostatistics, University of Kentucky, for help in biostatistical evaluation of the data Reprint requests to Dr H H Malluche, Division of Nephrology, Bone and Mineral Metabolism, Department of Medicine, University of Kentucky Medical Center, Lexington, Kentucky , UA References 1 HAUCHKA PV, LIAN J13, GALLO PM: Direct identification of the calcium binding amino acid-carboxyglutamate in mineralized tissue Proc Nail Acad ci LIA 72: , PRICE PA, OTUKA A, POER JW, KRITAPONI J, RAMIN N: Characterization of a 'y-carboxyglutamic acid containing protein from bone Proc Nat! Acad ci UA 73: , PRICE PA, NIHIMOTO K: Radioimmunoassay for thç vitamin K-dependent protein of bone and its discovery in plasma Proc NatI Acad ci UA 77: , DEFTO U, PARTHEMORE JG, PRICE PA: Changes in plasma bone Gla-protein during treatment of bone disease Calcif Tissue mt 34: , jovw DM, GUNDBERG OM, LIAN JB, NEER RM: Clinical evaluation of bone turnover by serum osteocalcin measurements (abstract) Ca/c,f Tissue mt 34:15, DELMA PD, WAHNER HW, MANN KG, RIGG BL: Assessment of bone turnover in postmenopausal osteoporosis by measurement of serum bone Gla-Protein J Lab Clin Med 12:47 476, DELMA PD, WILON DM, MANN KG, RIGG BL: Effect of renal function on plasma levels of bone Gla-protein J C/in Endocrinol Metab 57:128 13, PiucE PA, WILLIAMON MK, LOTI-IRINGER JW- Origin of the vitamin K-dependent bone protein found in plasma and its clearance by kidney and bone J Biol Chem 256: , CHEUNG AK, MANOLAGA5 C, CATHERWOOD BD, MO5ELY CA, MITA JA, BLANTZ RC, DEFTO U: Determinants of serum l,25(oh)2d levels in renal disease, Kidney bit 24:14 19, MALLUCHE HH: The value of bone biopsies for diagnosis of renal bone disease, in Dialysis Transplantation, Nephrology, edited by

6 874 Malluche et al MOORHEAD JR, BAILLOD PA, MION C, London, Pitman Medical, 1973, pp GOLDNER J: A modification of the Masson trichrome technique for routine laboratory purposes A,n J Pathol 14: , MALLUCHE HH, MEYER W, HERMAND, MARY G: Quantitative bone histology in 84 normal American subjects Micromorphometric analysis and evaluation of variance of iliac crest bone Calcif Tissue mt 34: , LILLIE PD, FULLMER HM: Histopathologic Technique and Practical Histochemistry (4th ed), New York, McGraw Hill, 1976, pp MALLUCFIE HH, HERMAN D, MEYER W, MANAKA R, MARY G: A new semiautomatic method for quantitative static and dynamic bone histology Calcif Tissue mt 34: , EGRE GV: Amino-terminal radioimmunoassays for human parathyroid hormone, in Clinical Disorders of Bone and Mineral Metabolism (International ngress eries 617), edited by FRAME B, POTT it, Amsterdam, Oxford, Princeton, Excerpta Medic 1983, pp BARR AG, GOODNIGHT GH, ALL GP, HELWIG GP: A User's Guide, Raleigh, A Institute Inc, 1979, pp MALLUCHE HH, FAUGERE MC: Histopathology of bane, in Renal Osteodvstrophy, edited by COBURN JW, CATTO GRD, Hingham, Martinus Nijhoff, in press 18 FARLEY JR, BAYLINK DJ: Purification of a skeletal growth factor from human bone Biochemistry 21: , NIHIMOTO K, PRICE PA: ecretion of Vitamin K-dependent protein of bone by rat osteosarcoma cells J Biol Chem 255: , PRICE PA, OTUKA A, POER JW, KRITAPNIs J, RAMAN N: Characterization of a y-carboxyglutamic acid-containing protein from bone Proc Nail Acad ci UA 73: , POER JW, PRICE PA: A method for decarboxylation of y-carboxyglutamic acid in proteins JBiol Chem 254: , PRICE PA, WILLIAMsON MK, HABA T, DELL RB, JEE W: Excessive mineralization with growth plate closure in rats on chronic warfarin Proc NatI Acad ci UA 79: , HAUCHKA PV, REDDI AH: rrelation of the appearance of y-carboxyglutamic acid with the onset of mineralization in developing endochondral bone Biochem Biophys Res mmun 92: , 198